Bottom Line

Gold has been my favorite investment since 2006. Since then, gold prices have risen more than 60% while the stock market has declined nearly 25%. My outlook for gold today is a bit more cautious than in 2006, but I still think opportunity for more upside exists. Thus, I am staying with gold for now.

Preview

This is the second of three newsletter issues that explore the concept of cycles as a long-term investing thesis. Last month, I studied stock market cycles. Below, I consider gold cycles. Next month, I will analyze real estate cycles.

Gold has been my favorite asset class since I began writing:

• Go for gold, October 2006
• A historical look at gold and the Dow, November 2006
• A testable prediction, April 2007
• Introduction, January 2008
• The case for buying gold, April 2008

I do not intend to hold gold forever. I would prefer to invest in stocks or real estate, but I don't see great value there. Eventually, I will trade my gold for these assets. But not yet.

I maintain a positive outlook on gold for several reasons: (1) gold is still relatively unpopular among investors, (2) gold's upward price momentum since 2001 is likely to continue, and (3) the length of the current gold bull market is still a bit short relative to history. More details below.

Why gold?

Below are some excerpts from my April 2008 newsletter, The Case for Buying Gold:

"The first step is to recognize gold as an asset class just like stocks, bonds, real estate, or commodities. And you should be willing to own some at the right price..."

"Long-term investors shouldn't care whether gold goes up in dollar terms, only whether gold is able to buy more stocks in the future..."

"Even though Bernanke is not *currently* increasing the money supply, he certainly can and I think ultimately he will. In Bernanke's famous Federal Reserve speech, Deflation: Making Sure 'It' Doesn't Happen Here, Bernanke makes it clear that he is prepared to run the printing presses to fight off deflation. He does not want to, but eventually he will have to. The political pressure to fight asset price deflation will be enormous..."

"The ultimate outlook for gold depends on how nasty you view the current credit crisis and how aggressive you think the Fed will act to try to resolve it. As for me, I think the crisis is about as big as they come and that the Fed will do anything and everything in their power to prevent it. And so I own gold."

I put the second quotation in bold since it reflects my overall attitude towards gold investing. I generally think of gold as a means of preserving wealth when more desirable investment classes are too expensive. Gold is good for this purpose since nobody can print more of it.

Suppose I think that stocks are expensive. I want to preserve wealth now so that I can buy more stocks in the future. However, I don't know if stocks will get cheaper in nominal terms (falling stock prices) or real terms only (adjusting for inflation). If I own gold, I don't care whether gold stays the same price and stocks get cheaper (deflation scenario) or stocks go up in nominal values but gold goes up even more (inflation scenario). Either way, gold buys more stocks over time.

If I hold cash, what prevents the government from inflating away my purchasing power? Not much, especially with a Federal Reserve determined to fight deflation at any cost. Owning gold removes this risk factor.

Long-term gold cycles

The other reason to own gold right now is that gold prices have been rising for 8 years since bottoming near $255 per ounce in April 2001. I believe this momentum will continue for another few years.

Some readers may be familiar with the "Dow to Gold Ratio", equal to the price of one share of the Dow Jones Industrial Average divided by the price of one ounce of gold. Consider this chart from an iTulip.com article from September 2001 that encouraged readers to invest in gold within months of the decade-long price bottom:

source: iTulip.com, "Questioning Fashionable Financial Advice", September 2001

I first became interested in gold based on a dow-to-gold ratio chart like the one above. Note the "Gold is Dead" and "Equities are Dead" labels, which arise from an investing cycle perspective.

Here is the ratio with data updated through October 10, 2009 (also graphed with my preferred index, the S&P 500, although you can see below that it doesn't make much difference):

Many followers of the dow-to-gold ratio expect a return to parity - that is, one ounce of gold will eventually buy one share of the Dow Jones Industrial Average. After all, the previous two minimums occurred at 1.74 in 1932 and 1.27 in 1980. I understand where these investors are coming from, but I think that holding out for parity could be an error. Not only is it a better strategy to scale out gradually as the extreme valuation approaches, but I'm not convinced the ratio will go that low this time around.

Consider that the peaks in the dow-to-gold ratio are growing over time. Also consider what the trend looks like when graphed back to 1870:

In my view, the only reason why the first chart looks "flat" is because of the extreme overvaluation of gold in the 1980s - a true gold bubble. There is no economic reason why the stock market and the price of gold should have the same long-term growth rates. It is unsurprising to me that in the long-run, stock prices increase faster than gold prices. The truth probably lies somewhere between the two representations above, though closer to the second.

In any case, investing/valuation cycles are evident in both charts. Gold and stock valuations have tended to move from one extreme to the other over multi-year intervals. Similar to the Bull/Bear table from last month's newsletter, here is a table of stocks versus gold outperformance over time:

Stocks outperform gold in each of the "Stock" periods, and gold outperforms stocks in each of the "Gold" periods. No magic number of years determines the length of these time periods, but stock bull markets tend to last longer than gold bull markets.

If you attribute all of the variation in the table above to fluctuations in the stock market, you are missing a key point: that gold - not dollars - provides the best benchmark to measure stock market value. Consider that the stock market rose 20% in dollar terms between 1967 and 1980, but its purchasing power went down over 90% when measured in gold.

The way I see it, gold is no longer a great value relative to stocks. In fact, I would say we are closer to the end of the gold cycle than the beginning. The stock-to-gold ratio approaching the lower red trendline certainly warrants caution. Still, it doesn't feel like a gold top. Even though investors are slowly coming on board, most people I know don't own any.

Consider the following excerpt from Minyanville:

I just love it when I run across an article like the one I just read in the Wall Street Journal this week.

The headline: Gold Is Still a Lousy Investment

I read it differently, as: Gold Still Has a Long Way to Go

You know you’ve reached the top in a bull market when the shoeshine boys and taxi drivers who heard a hot tip are clamoring to get in. Two years ago, when people in cocktail parties in Florida were talking about how you can’t go wrong in Miami real estate, you knew it was time to sell.

- The Cool Kids Still Like Gold, October 9, 2009

I agree. Here's how I would sum up my gold investment stance: make sure you own some, but don't go all-in. You should be willing to reduce exposure over time if the price keeps rising. I will probably begin to scale out of gold if the price rises into and beyond the $1200-1500 range. The great value years are over, but the momentum is likely to continue. We are not in a gold bubble, so downside risks are not extreme. All things considered, I still like gold for the next few years.

Back to cycles

Last month's newsletter generated some interesting questions from a thoughtful reader. He writes:

[Cycle investing] seems like a convenient framework, but it's unsettling to me for a number of reasons. How long does a cycle last? The traditional answer is "well, there are short-term cycles and long-term cycles." The thesis of your article, which makes a lot of sense, is that the real prudent investors will consider the long-term cycles. But if you're willing to actually keep an eye on your investments (which has been one of the points that you've made fairly frequently in your relatively unsympathetic (and perhaps rightly so) view of the plight of the unsophisticated investor), shouldn't the real opportunities for higher returns be in these short-term cycles?

I think that opportunities to take advantage of asset mispricings occur over many time horizons. Do I believe that overvaluation/undervaluation occurs in day trading? Yes. Do I believe that overvaluation/undervaluation occurs over weekly/monthly timeframes? Yes. Do I believe that overvaluation/undervaluation occurs over longer, multi-year horizons? Clearly yes.

But I also find that betting on shorter cycles is more difficult and less fruitful than betting on longer cycles. If you ask me to tell you whether the stock market will be higher or lower in two weeks, I have no idea. But if you ask me whether the stock market will be higher or lower in two years, I am more willing to make a bet. Some traders are successful over the short-run, but it takes tremendous time and effort. It is no coincidence that legendary investors like Warren Buffett and Jim Rogers have the most success over longer time horizons. Long-run asset mispricings are easier to take advantage of because you do the analysis once and then stick with the same plan for years.

He continues:

"Taking the existence of cycles as a given" also is unsettling because it seems like the Texas sharpshooter problem to me. Even if there is a cycle, now you're just trying to fit today's data into the context of that cycle, not knowing how large the amplitude of a fluctuation or whether you're in a mini-peak or a larger peak."

This is a legitimate concern, but I don't think this fallacy applies to my formulation. My investing analysis is far from a data-mining exercise. Consider how I go about the analysis. Step one: propose guiding principles that arise from experience, reading, and broad interaction with the data:

(1) Assets can be overvalued or undervalued (prices do not always reflect fundamental value)
(2) Overvaluation/undervaluation can last for years (e.g. Internet Bubble, Housing Bubble)
(3) Undervalued extremes tend to follow overvalued extremes, and vice versa

All of these are grounded in a basic understanding of markets, economics, and human nature. Together, these principles make up the model/hypothesis. Step two: use data to calibrate the model (i.e., identify multi-year highs and lows, compute the time between them, study the time periods, etc.). Step three: using insight from the model and calibration from the data, I make out-of-sample predictions for the future. This overarching theory explains why I have been bullish on gold and bearish on stocks since 2006.

It seems like you've adopted a fifteen-year length of the cycle approach. But why is that? You also stated that you tended to agree that "after the 2000-2002 and 2007-2009 recession, one more should be enough." Why is that? Anecdotal studies of investor confidence?

For cycle length, it begins with a recognition that bubbles occur over multi-year time frames. How long does it take for bubbles to recover? The Japanese Nikkei is currently 75% lower than its peak in 1989. The Nasdaq is nearly 60% lower than its peak in 2000. In other words, if you're not looking at a multi-year timeframe, you're missing the biggest price booms and busts.

Intuition about human behavior also supports longer term investing cycles. It takes years for industries to change and for people to forget investment trends. Consider how many are expecting the Housing Bubble to resume a V-shape recovery.

After accepting a multi-year timeframe, I let the data guide the exact calibration. Historically, stock/gold cycles tend to be 10-20 years in length.

Finally, I always wondered about how the specific commonly cited causes of the recent crashes play into the overall "there must be cycles" paradigm. I know that you blame the housing bubble on the Internet bubble, so let's start at the beginning of this cycle in 2001, when the "Internet bubble burst." In past bear markets, what was the "bubble"? According to the "cycle" theory, there would have been a long bear period even without the bursting of the Internet bubble. So can the Internet bubble be better characterized as "after the high-point in the cycle, investors were doing so well that they started making stupid investments in unprofitable businesses? If so, or if there's another explanation, can you go back in time and describe how the other events that kicked off the troughs of the long-term cycles fit into that explanation? I really want to believe in the "cycle theory," but there has to be a more scientific (or even more anecdotal) way of explaining it.

All of the major stock market peaks in the last century have stories and legacies behind them: the Dot-com Bubble of 2000, the Go-Go Years of the 1960s, the Roaring Twenties and the Wall Street Crash of 1929, and finally the Panic of 1901, which I believe occurred as a result of overvalued railroad stocks. Consider also the stories behind the Florida land boom of the 1920s and the most recent U.S. Housing Bubble. For a more complete analysis of major market bottoms, see the May 2009 newsletter: Bear market anatomy.

In all of these cases, investors were chasing a story - an intuitive and believable narrative - for why prices tomorrow would be higher than prices today. Consider today's story for gold: the economy is crashing, congress is running huge deficits, the Fed is printing trillions, inflation is inevitable, blah blah blah. All of them are true, of course - I've written about them in the past - but the stories ignore price. Does the price of gold already reflect these factors or not? Smart investors know that price is everything.

With regard to the Internet Bubble specifically: yes, I think that stocks would have entered a long term bear market even if they had peaked out in 1996 prior to the extremely speculative years from 1996 to 2000. In my mind, the word "bubble" characterizes only the last few years of a bull market where prices are completely removed from fundamentals and increase only from momentum.

As for more scientific approaches to studying bubbles, trust me that the academic literature in behavioral finance is huge. New theories are developed and tested all the time. I should know, given that I am a recent product of Bernanke's Bubble Laboratory. Try the homepage of Princeton's Harrison Hong for a sampling of what's out there. Some good, some bad, but almost always interesting.

Conclusion

Circling back, consider the life cycle of a representative asset. At major market lows, investors can buy the asset for less than fundamental value. Sophisticated investors tell their friends to buy, but no one listens. Eventually, enough investors recognize the asset as cheap so that prices start to rise. The sophisticated investors look smart in retrospect, and new investors pile in over time. The investing public is convinced slowly as prices rise and rise. At some point, rising prices becomes the sole basis for investing. Market tops occur when every investor who will invest has already invested. No one is left to drive the price higher. And then prices start to fall...

Consider this nice visual representation of the gold market from the folks at iTulip.com. I've learned over the years to take their advice seriously (take a moment to review "iTulip's Record" at the top of the page).

source: iTulip.com

Gold is tough to value since typical metrics like price to earnings multiples or price to rent ratios do not apply. As such, an investing cycle framework is particularly helpful in this case. As for me, I'm not selling my gold...not yet at least.